About

Almost two decades of fast, accurate numerical software.

Every new rule raises the compute bill: tens of thousands of sensitivities, every book, every day. The code banks have already validated should keep up, without being rewritten.

Our story

One purpose since our first client in 2010: help banks keep up with ever tougher regulation, from our first high-performance software to XAD and its code generation backend.

  1. 2006

    Live signal processing

    Our founders meet at Trinity College Dublin, working on one problem: making compute-intensive applications run in real time. The field is telecoms, live signal processing, where a late answer is no answer.
  2. 2009

    Finance after the crisis

    The research matures, and finance comes knocking. After the 2008 crisis, regulators tighten how banks measure risk: XVA arrives, and Monte Carlo simulations need many times more scenarios. The same problem, with far higher stakes.
  3. 2010

    First client: a European bank

    A European bank becomes our first client, and Xcelerit is spun out of Trinity College Dublin. The Xcelerit SDK follows, and spreads through bank pricing and risk code.
  4. 2012

    The benchmarks begin

    Our first openly published finance benchmarks: processors and accelerators on the same workloads, shared with the community.
  5. 2014

    From faster valuations to faster sensitivities

    Sensitivity-based rules, on the road to FRTB, SIMM and SA-CVA, turn the problem into sensitivities. Even after making a valuation ten times faster, ten thousand sensitivities with bumping still cost a thousand original valuations, too many to finish overnight. So we turned our know-how on the sensitivities themselves: QuantAD, our commercial AAD product, gives every one of them for the cost of a handful of valuations instead of one per risk factor.
  6. 2022

    XAD, in the open

    QuantAD becomes XAD and goes open source, with QuantLib integration from the first release. Python and QuantLib-Risks follow in 2024.
  7. 2026

    270× on QuantLib

    XAD code generation goes further. QuantLib with AAD computes all sensitivities of a pricing up to 270× faster than bumping on Monte Carlo, tested daily against QuantLib's master. XAD internals presented at CppCon.

Along the way: two-time winner of HPCwire's award for the best use of HPC in financial services; PRACE's award for the most innovative industrial HPC application in Europe; and Red Herring Top 100 finalist, Europe and Global.

The people behind Xcelerit

A spin-out of Trinity College Dublin's telecoms research centre, CTVR, now CONNECT.

Leadership
Hicham Lahlou

Hicham Lahlou

Co-founder and CEO

Researcher at Trinity College Dublin's CTVR telecoms research centre before co-founding Xcelerit. INSA Lyon engineer, first in space-sector software. Leads the company and its work with banks.

LinkedIn
Jörg Lotze

Jörg Lotze

Co-founder and CTO

PhD at Trinity College Dublin, after engineering roles at Xilinx and Melexis. Architect of XAD; presented its internals at CppCon 2026.

LinkedIn
The team

A Dublin team of XAD core developers, quant developers and C++ performance engineers. You work directly with the engineers who build the code.

Clients
  • Deutsche Bank
  • Société Générale
  • Nomura
  • Lloyds Bank
  • Intesa Sanpaolo
  • European Investment Bank
  • European Bank for Reconstruction and Development
  • Aegon
  • aixigo

Related company: Norma Analytics, live risk analytics for derivatives.

Join the team

We hire engineers who like hard numerical problems and shipping the answer to production.

Open roles